A network psychometric analysis of teachers' AI adoption profiles and occupational perceptions

Teachers' responses to artificial intelligence (AI) are mostly studied through intention‐based acceptance models, where a binary contrast between users and non‐users leaves the differences inside each group out of view. Occupational variables, when considered, are typically examined one at a time. This study applied diffusion of innovations theory to move beyond the user/non‐user dichotomy towards empirically derived adoption profiles, and modelled occupational perceptions as an interconnected network instead of as separate variables. Data from the TALIS 2024 AI module were analysed for 54,276 teachers across 55 education systems. Latent class analysis of teachers' AI use, attitudes, and barriers identified five distinct profiles corresponding to the theory's adopter categories. A mixed graphical model estimated the unique association between profile membership and each of 12 occupational perceptions. Perceived societal value of the profession emerged as the strongest correlate (edge weight = 0.24), three to five times larger than any other; self‐efficacy played a secondary role. Perceived value tracked openness to AI rather than actual use, while indecision and principled rejection concentrated among those feeling undervalued. Value was highest among teachers who had internalised AI pedagogically, not merely used it instrumentally. Beyond self‐efficacy and well‐being, these findings position the perceived standing of the teaching profession as a central condition for AI adoption. Policies that address how the profession is valued may therefore be as consequential for integration as those targeting skills or infrastructure.